Elixir Sulfanilamide: The 1937 Poisoning That Created the FDA — and the Glycol Mix-Up That Still Kills

In the autumn of 1937, a medicine that tasted of raspberries killed at least 105 Americans, 34 of them children — and the federal government’s only legal grounds for pulling it off the shelves was a spelling technicality. The active drug was fine. The poison was the solvent: a sweet, syrupy industrial liquid called diethylene glycol that nobody had been required to test, because in 1937 no law required a drug to be proven safe before it was sold. What happened next created the FDA as the world knows it. And the specific chemical confusion at the heart of it — one glycol swapped for another — has never stopped killing people. It killed more than 300 children as recently as 2022.

What was Elixir Sulfanilamide?

Elixir Sulfanilamide was a liquid preparation of sulfanilamide — one of the first genuinely effective antibacterial drugs — dissolved in diethylene glycol, flavored with raspberry extract, saccharin and caramel, and sold by the S.E. Massengill Company of Bristol, Tennessee, beginning in September 1937.

Sulfanilamide itself was a legitimate medical breakthrough. The first of the sulfa drugs, it actually cured streptococcal infections — strep throat, scarlet fever complications, the childbed fever that had killed mothers for centuries — in an era when almost nothing else did. Doctors prescribed it enthusiastically, in tablets and capsules, and it saved lives. Penicillin was still years from the pharmacy shelf; sulfanilamide was the wonder drug of the 1930s.

The trouble began with a marketing request. Massengill’s salesmen reported demand, particularly in the South, for the drug in liquid form — easier for children to swallow, easier to flavor, familiar to patients raised on cough syrups and tonics. The company’s chief chemist and pharmacist, Harold Cole Watkins, went looking for a solvent, because sulfanilamide dissolves poorly in water and in alcohol. He found that it dissolved beautifully in diethylene glycol. He added raspberry extract, saccharin and caramel, the lab checked the mixture for appearance, flavor and fragrance — and shipped it.

Nobody tested it for safety. Under the Pure Food and Drugs Act of 1906, the law then in force, nobody had to.

Why did Elixir Sulfanilamide kill 105 people?

Because its solvent, diethylene glycol, is metabolized by the human body into an acid that destroys the kidneys — and the elixir’s victims swallowed it by the spoonful, on their doctors’ instructions, for days.

Diethylene glycol tastes sweet and pours like syrup, which made it seem an ideal vehicle for a children’s medicine. Inside the body, the same enzymes that process ordinary alcohol go to work on it, oxidizing it stepwise into 2-hydroxyethoxyacetic acid. That metabolite accumulates in the kidneys and shuts them down. Victims typically suffered nausea, then stopped producing urine, then died of kidney failure over several agonizing days. There was no antidote in 1937, and dialysis did not exist.

By early October, doctors in Tulsa, Oklahoma, were reporting a cluster of unexplained deaths among patients who had all taken the new elixir, and the American Medical Association began receiving reports of fatalities. The federal Food and Drug Administration — then a small agency with almost no pre-market authority — threw essentially its entire field force at the problem: 239 inspectors and chemists chased 240 gallons of elixir across 633 shipments, working pharmacy ledgers, prescription pads and small-town gossip to find every bottle. They recovered 234 gallons and 1 pint. The missing few gallons — much of it already swallowed — accounted for the deaths.

A mother in Tulsa named Maise Nidiffer, whose six-year-old daughter Joan died after taking the elixir for a sore throat, wrote to President Roosevelt: “The first time I ever had occasion to call in a doctor for [Joan] and she was given Elixir of Sulfanilamide. All that is left to us is the caring for her little grave…” Her letter, and the photograph of Joan she enclosed, circulated through the press and Congress and did as much as any hearing to change American law.

The company’s founder, Dr. Samuel Evans Massengill, gave the response that became infamous: “My chemists and I deeply regret the fatal results, but there was no error in the manufacture of the product.” Within the letter of 1937 law, he was right — which was precisely the problem. Watkins, the chemist who chose the solvent, died by suicide in January 1939 while awaiting trial.

Diethylene glycol vs. propylene glycol: what is the difference?

They are near-twins on the shelf — both colorless, both syrupy, both faintly sweet — and the entire difference between a safe medicine and a mass poisoning lives in their structure and in what the body turns them into.

Propylene glycol (propane-1,2-diol, CAS 57-55-6) is a small three-carbon molecule carrying two hydroxyl groups. The body treats it almost like food: the same alcohol-processing enzymes oxidize it to lactaldehyde and then to lactate and pyruvate — molecules your muscles produce every time you exercise, fed straight into ordinary metabolism. That benign fate is why the FDA lists propylene glycol as GRAS (generally recognized as safe) for use in food under 21 CFR 184.1666, and why it became the standard solvent in liquid medicines, the carrier in theatrical fog fluid, and the working fluid in food-plant heat-transfer loops.

Diethylene glycol (CAS 111-46-6) looks almost identical in the drum and on a spec sheet’s physical rows — but structurally it is two ethylene glycol units joined through an ether oxygen, and the body’s same enzymes oxidize it to 2-hydroxyethoxyacetic acid, a metabolite the kidneys cannot clear. The estimated human lethal dose is in the range of a single mL per kilogram of body weight. A few spoonfuls a day for a week was enough to kill a child in 1937, and it still is.

There is a larger lesson buried in that table, and 1937 is when pharmacology learned it: the vehicle is a drug too. Before the elixir, an “inactive ingredient” was treated as scenery — the flavoring, the filler, the liquid that carried the medicine. Elixir Sulfanilamide is the case that taught regulators, chemists and formulators that every gram in the bottle goes into the patient, active or not, and that a solvent chosen for solubility and mouthfeel can outweigh the drug it carries. Modern excipient science — the entire discipline of qualifying the other ingredients — descends from those 105 deaths.

How did the FDA stop the elixir — with a spelling technicality?

Under the 1906 law, selling an untested poison as medicine was not illegal — but calling it an “elixir” when it contained no alcohol was. That labeling technicality, misbranding, was the only legal hook the government had to seize the product.

An “elixir,” in the pharmacopeial sense, is an alcoholic solution. Watkins had dissolved his sulfanilamide in diethylene glycol, not ethanol — so the product was, in the law’s narrow eyes, mislabeled. Had Massengill called the bottle a “solution,” the FDA would have had no authority to touch it while it killed. Agency historians have made the point plainly ever since: the federal government seized the deadliest medicine of the century on a word.

The S.E. Massengill Company ultimately pleaded guilty to 112 counts of adulteration and misbranding and paid a fine of $26,100 — the largest ever levied under the 1906 Act, and about $250 per death. That was the entire legal consequence available.

Why was the FDA created as we know it today?

The modern FDA — the agency that reviews a drug’s safety before it can be sold — was created by the Federal Food, Drug, and Cosmetic Act of 1938, and the 1938 Act was passed because of Elixir Sulfanilamide.

A stronger food-and-drug bill had been stalled in Congress for five years; the industry lobby was effective and the public was indifferent. The elixir deaths ended the indifference in a single news cycle. In June 1938, Congress passed the FDCA and Roosevelt signed it. The new law required manufacturers to submit a new drug application demonstrating safety before marketing, gave the FDA authority over cosmetics and medical devices, required adequate directions for use, and abolished the loophole that had made the elixir legal until it was misspelled.

One boundary is worth drawing precisely, because it is where most retellings go wrong: the 1938 Act required proof of safety, not proof that a drug worked. Efficacy requirements came only in 1962, with the Kefauver-Harris amendments — passed after a different near-disaster, and thanks in large part to a scientist whose career the elixir had launched.

Who was Frances Oldham Kelsey?

Frances Oldham Kelsey was the FDA reviewer who refused to approve thalidomide for the American market in 1960–62 — and her first professional encounter with a drug disaster was Elixir Sulfanilamide, as a young researcher in the University of Chicago pharmacology laboratory that helped confirm diethylene glycol was the killer.

In 1937 the FDA needed to know, with courtroom certainty, which ingredient in the elixir had done the killing — the drug, the flavoring or the solvent. Animal studies run in E.M.K. Geiling’s laboratory at Chicago, where the young Frances Oldham was then working, established that the diethylene glycol was responsible. Twenty-three years later, as a new medical reviewer at the FDA, Kelsey sat on the application for thalidomide — a sedative already sold across Europe — and kept asking the manufacturer for the safety data the 1938 law entitled her to demand. The data never satisfied her; she never approved it; and when thalidomide’s catastrophic effects on developing limbs surfaced in Europe, the United States had been spared almost entirely. President Kennedy gave her the President’s Award for Distinguished Federal Civilian Service in 1962.

Her career is the two great drug-safety laws of the twentieth century, bookended: the 1938 Act that the sulfanilamide deaths wrote, and the 1962 amendments that her thalidomide stand made possible. It began with a glycol.

Is diethylene glycol still poisoning people today?

Yes — the identical substitution that killed 105 Americans in 1937 killed more than 300 children in The Gambia, Indonesia and Uzbekistan in 2022–2023, in cough syrups where diethylene glycol or ethylene glycol had taken the place of pharmaceutical-grade propylene glycol or glycerin.

Beginning in September 2022, clusters of acute kidney injury in small children — the same presentation Tulsa doctors saw in 1937 — were traced to imported syrups contaminated with DEG and EG. The World Health Organization issued five separate medical product alerts through mid-2023, naming syrups manufactured in India, and counted roughly 70 deaths in The Gambia, more than 200 in Indonesia and about 20 in Uzbekistan, the great majority children under five. Further deaths linked to a contaminated syrup were reported in India itself in late 2025.

Nor was 2022 the first echo. In Haiti in 1995–96, dozens of children died of kidney failure from an acetaminophen syrup made with DEG-contaminated glycerin — the other sweet polyol that shares this failure mode, covered in our glycerin formulation guide. In Panama in 2006, DEG sold as glycerin found its way into government-produced cough syrup and killed at least dozens, with credible estimates running far higher. The Panama investigation traced the poison backward through a chain of brokers and relabeled drums across three continents — at each hand-off the paperwork got friendlier and the origin got vaguer, until barrels of industrial diethylene glycol wore the documentation of pharmaceutical glycerin. Nobody in the chain had tested the contents. Every link had trusted the label written by the link before.

The economics have not changed since Bristol, Tennessee: DEG is cheap, looks and handles exactly like the glycols that belong in a syrup, and enters the supply chain wherever a buyer takes a drum’s label on faith. In response, the FDA’s current guidance on high-risk drug components — glycerin, propylene glycol and other polyol ingredients — tells manufacturers to identity-test every container of every lot for DEG and EG before use, against a safety limit of not more than 0.10%. The United States Pharmacopeia’s propylene glycol monograph carries the corresponding limit tests. Eighty-nine years of regulation, distilled: never trust the label on a sweet, syrupy, colorless liquid — test it.

How do you know a glycol is actually what the label says?

You know because somebody ran the analysis on the actual lot in the actual container — and handed you the paper. That is what a certificate of analysis is for, and the entire history above is the argument for reading it.

• Buy the grade the application requires. If a glycol touches food, drink, medicine or anything people ingest or absorb, it should be USP grade — material tested against the pharmacopeial monograph, including the DEG/EG limit tests. Technical grade belongs in heat-transfer loops, deicing, concrete curing and industrial formulation, where the specification is fitness for process, not fitness for people.
• Match the fluid to the loop. For chillers, hydronic loops and process cooling, an inhibited propylene glycol — PG with a corrosion-inhibitor package — protects the metal in the system the way bare glycol cannot. Our brewery chiller glycol guide covers the concentration math.
• Insist on a certificate of analysis per lot — not a generic spec sheet. A CoA states what the analysis of that batch actually found. Every Alliance Chemical order ships with one.
• Know your supplier’s supply chain. The 2022 poisonings happened at the ingredient-broker layer, where drums change hands and paperwork gets friendlier at each step. Shorter, documented supply chains are the structural fix.

What identity testing actually means — and what it does not

It is worth being concrete about the word “test,” because the failure mode in every one of these poisonings was a check that felt like testing but was not. Appearance is not a test: PG and DEG are both water-clear. Density and refractive index barely separate them, and never in a mixture. Taste — the 1937 lab’s actual method — is the poison working. A photocopied certificate traveling with a relabeled drum is not a test; it is the vector. The checks that actually distinguish one glycol from another are instrumental: an infrared spectrum or a gas-chromatography run compared against the monograph standard, performed on a sample drawn from the container in front of you. That last clause is the heart of the FDA’s current rule — every container, not a composite of the lot, because contamination travels drum by drum, not lot by lot.

Key numbers and sources

Related reading

• An introduction to propylene glycol — the working reference for PG itself: properties, grades and applications.
• Is propylene glycol safe? — the modern safety question, answered with the regulatory record.
• Propylene glycol vs. ethylene glycol — the other glycol confusion, and when each belongs in a loop.
• PG in food and beverage manufacturing — where USP grade earns its premium.
• Best glycol for brewery chillers — concentration math for real systems.
• Who discovered phosphorus? — Hennig Brand, 1669, and chemistry’s strangest discovery story.
• Ignaz Semmelweis and chlorinated lime — the chemistry that proved handwashing worked, decades before anyone believed it.